Modified p-median approach for efficient GT cell formation

The objective of this paper is to propose the modified formulations over the classical p-median model for efficiently solving the cell formation (CF) problem with which production managers are faced in implementing group technology (GT) to manufacturing. In spite of its merits as the robust solution approach to GT cell formation, the classical p-median model is limited to small-size CF problems since it requires many binary variables. Necessity of many binary variables makes it difficult for cell designer to use the classical p-median model for solving even medium-sized CF problems, let alone large-size CF problems even though large-memory main frame computer and highly expensive commercial software are used. To resolve this sort of problems, we propose two modified p-median formulations which are even applicable to largest CF problems among the ones found in literature. Their main virtues are as follows. First, the modified p-median formulations encourage speedy implementation of the models. Secondly, the proposed formulations that are applied to moderately large-size CF problems with 30 or more machines can run effectively even on personal computer and education-purpose softwares such as the HYPER LINDO which are available at such a low cost compared with commercial software. Computational results from applying to moderately large-size CF problems with 30 to 50 machines show that the CF using the modified p-median formulations provides very promising implications as compared with the CF using the classical p-median model.